This proposal is designed to study the social relationships of physically abused children of school age. Recent work indicates that abused children have lower social status among their peers than non-abused children and that their relationships with peers are characterized by less positive reciprocity and more negativity. Such findings suggest that these children are at seriously increased risk of diminished social and emotional support as they grow older and move into adolescence. The study addresses the questions of how their diminished social status comes about, whether their conceptions of friendship and understanding of social support differ from those of non-abused children, and what the effects are of lower social status on the children's feelings about their own social relationships. It proposes a set of three models to account for the effect of abuse on sociometric status, quality of friendship, and feelings of loneliness and social dissatisfaction. The first two models examine how two mediating factors (children's social cognition and their social behavior with peers) bring about lower social status and lowered expectations for friendship. The third model examines the effects on loneliness and social dissatisfaction that are associated with lowered sociometric status and poor quality of friendships. One hundred physically abused children are recruited from the confirmed cases of abuse on the NYC Special Services for Children Register and compared with 100 non-abused classmates matched case by case for gender, age, race, ethnicity, and SES. Sociometric assessments will be conducted in classes, and individual interviews will be conducted elsewhere in school. Teachers and parents will be asked to rate the children for social competence and problem behavior, and mothers will be interviewed at home to obtain demographic information about the family, parenting practices, and descriptions of interactive behavior among family members. Peer relationships should be of special importance for abused children in light in their highly stressful family relationships. This may well constitute a domain in which early intervention would have unusually good chances of success, since studies of social/behavioral intervention with children low sociometric status have shown promising results. The data provided by this study would be valuable in the design of such intervention.